Navigating Public Risk with LLMs: A Few-Shot Analysis of Social Media Sentiment and Commun

Harvard Dataverse hosts the dataset and code for the paper 'Navigating Public Risk with LLMs: A Few-Shot Analysis of Social Media Sentiment and Communication Strategies in the Nuclear Wastewater Discharge Crisis'. The repository contains raw data and experimental code for comparative and main analyses. The dataset was last updated on April 20, 2026.

Use Cases

• Analyze public sentiment on social media regarding the nuclear wastewater discharge crisis based on the dataset's description.
• Compare communication strategies used during the crisis based on the provided experimental code.
• Conduct few-shot learning experiments with LLMs for sentiment classification as referenced in the paper title.
• Study the relationship between social media discourse and public risk perception in environmental crises.

Strengths

• Dataset is directly linked to a specific, named research paper, providing clear academic context.
• Includes associated code for both comparative and main experiments, facilitating reproducibility.
• Last update timestamp is precisely recorded as 2026-04-20 14:01:13.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Harvard Dataverse, authored by he, lingfei.

Collection Method

Likely contains social media data collected for the associated research paper.

Time Range

null

Freshness

Last updated 2026-04 20 14:01:13; freshness should be verified.

Geography

null